Fix PR status fail for large builds (#2290)

### ❔What, Why & How

Correctly return data when getting the PR status for builds with over a
hundred jobs.

`--slurp` ensures a single object, instead of one object per 100 jobs.

Also limited the call to 50 per page, to lower request time per call
which minimized chance of 502 error. Added retry logic as well.

Related to issue: #

### ✅ Checklist

- [x] Add tests (E2E, unit tests)
- [x] Update RELEASENOTES.md
- [ ] Update documentation (e.g. for new settings or scenarios)
- [ ] Add telemetry

<!-- Include more checklist entries, if needed -->

---------

Co-authored-by: spetersenms <spetersenms@users.noreply.github.com>
Co-authored-by: Alexander Holstrup <117829001+aholstrup1@users.noreply.github.com>